A letter to my students about why I disagree with Paul Kirschner about the Failure of Constructivist/PBL/IBL etc…

Context – my students learning to consider the viability of Learning Styles

Had a great time with my ed tech pre-service students last week. We were learning about searching (the web) and research and talking about how to be effective in learning what we need to know about tools and approaches that we come across. My edtech courses have never been particularly content focused, but after some discussions with Tom Farrelly this summer, I’ve converted it almost entirely to teaching the literacies I think students need to discover what they need to know based on their own values.

The first hour of the class was focused on looking at the learning styles information on the web and comparing it to existing research. The vast majority of my students come into class believing in learning styles and, for many, it’s the only educational theory language they are comfortable using. Our first group read was “The myth of learning styles” by Reiner and Willingham. My purpose in choosing that particular article is that, while I tend to agree that the concept of learning styles has serious limitations, at least in Willingham’s case, I don’t tend to be on… his side of education. He is more invested in memory than I am and thinks that expert learners are people like chess players. He’s a huge figure in cognitivist literature dealing with education. I used him because I wanted students to understand that research comes from a context, and finding out about that context can help you understand what a person means by words like learning.

Look at the intersection of memory and chess. The ability to remember every pattern on a chess board is going to be hugely important for people trying to be good at chess. While there are LOTS and LOTS of potential patterns, there are a limited number, and chess has clear rules about winning and losing. Much like the other examples Willingham uses, like computer science and music, we can understand how memory is going to be hugely beneficial to people working at a high expert level in those fields.

That’s not me. It’s never going to be me. I’m never going to be a world class expert in a field like chess or computer science. I would also argue that almost none of my students will be either. The important question, I think, is to consider what things we do value preparing ourselves for and considering whether our approach to teaching best prepares them to do that.

So when I offhandedly suggested to my students that there is a whole field of education that is committed to that kind of work, they quite rightly asked for the research. 🙂 A review of that literature is out of the scope of the course I’m teaching, so on the off chance that some of them are interested, I thought i would put a quick article breakdown here of one of those pieces of work.

Why Minimal Guidance During Instruction Does Not Work: An Analysis of the Failure of Constructivist, Discovery, Problem-Based, Experiential, and Inquiry-Based Teaching

This article has been cited over 10000 times according to google scholar, so safe to call it influential. I’m using it here because it allows me to point to some of the patterns I’ve blogged about here before, all wrapped up in a nice tight package. If you’re a constructivist, I encourage you to read the article, I think it’s a nice introduction to what the constructivist haters think. If you love Kirschner et. al., I’m happy to engage in a conversation, but it will have to start with our lack of shared epistemology. I find that conversation fascinating, I don’t find the ‘but science!’ conversation to be as fascinating.

Minimal guidance

Here’s an easy one to start with. The citation provided tying constructivist pedagogies to ‘unguided’ or ‘minimal guidance’ are not, I would suggest, sufficient to the use of the term. My constructivist classrooms are VERY guided though it’s true that I’m not handing over tons of content for people to memorize. I see the expression ‘unguided’ or ‘minimal guidance’ as a misrepresentation of what constructivism is about.

A good reference here is the Mayer article (firewalled) cited throughout. The article, from 2004, suggests that totally unguided instruction is not the best way to have people accomplish defined tasks. Totally agree. I have never found a problem based teacher, for instance, who is providing ‘totally unguided instruction’. Sure. People aren’t going to discover an algorithm for solving a math equation by just hanging around some numbers. Agreed.

The chess argument

I’ve made a few comments about chess above, so I wont go over it again. Simply put, chess is a game that you can win. Most of the important decisions people make in their lives aren’t games with rules that show how to win. I think that makes it a suspect example for helping people become learned humans. Same for solving math problems. Same for computer programming. They are important niche skills, certainly, I’m not convinced they are the basis for learning as a human. Nathan Ensmenger’s article on chess in education is excellent.

Problem solving skills

Much of the cognitive argument in the article is around what works best for ‘problem-solving skills’. This is where your values come into play. Are we, on the whole, teaching ‘problem solving skills’ in our education system? You might see it that way. When I look at the regular life of a regular person there are certainly problems that are solved. I can fix a leaky tap. I can, to use the example in the article, cross a road without being hit or do a math problem. Most of the things I decide on in a given day, however, are not these kinds of things. Not in my job. Not in my family life. Not doing construction in my attic. There’s no one to tell me I’ve done the work right. I am always choosing between a variety of intersecting and often conflicting rules, values and implications.

For me, constructivism is not about teaching people to problem solve the MOST EFFECTIVELY. It’s about learning to confront uncertainty. It’s about learning to ask a question even when it’s not clear what that question should be. It’s about deciding when you don’t have all the information. It’s preparation for life.

So I don’t consider constructivism a failure if its not the best preparation for problem solving. Solving (getting the right answer) problems is not the thing on the top of my list in my classroom.

Novices and experts

This one is trickier and one I struggle to explain. The argument the article seems to make is that because novices don’t have loads of information in long term memory, it’s harder for them to use that information to do the work. Putting aside that ‘the work’ that the article wants us to do is solve problems with right answers, I have some other concerns.

Most of us will always be novices at almost everything. The pathway that seems to be implied here is – master the content – then you can use the higher order thinking as an expert. But most of us will never be experts at what we’re learning. Most of my students will be english teachers or math teachers but not ‘edtech experts’. If I give them information now, most of them will never do enough work to become an edtech expert to allow them access to the higher order conversation. I’ll be preparing them with information from 3-5 years ago, for a career 3-5 years in the future. Information that will continue to be out of date as we go forward.

My view of constructivism reaches for a modified guided expert approach. Sure. If I say ‘hey, go out and evaluate this math software for the classroom’ with no support, it’s going to be terrible for them. But going through these guided approaches, where I spot them a few questions, do a lot of iterative feedback, and allow them to develop their skills, gives them the chance to get some of those expertish tools that can help them later.

Final notes

The upshot of my concern with this research is that it reduces teaching to ‘helping people solve problems’. It also sets up a hierarchy of learning where people get basic instruction now and expert instruction later, even though the vast majority of us never make it to the later part. I’m not interested in populating the world with more excellent chess players, I’m more interested with compassionate citizens who can engage in difficult discussions in ways that help us work through the challenges in our society.

Teaching with AI checklist

I pulled together my notes from the last 9 months of doing sessions on Generative AI and compiled them yesterday. Shout out to Nick Baker for adding some things, editing some things and overall making it more nicely to readish.

Every time I approach this issue I keep thinking that so much more and probably a lot less should be said. Every time I meet with a group of faculty or teachers on this issue we go through a few phases

  1. Boredom, kinda, as I explain what generative AI is.
  2. A bit of ‘yeah, this doesn’t apply to me, my courses…’
  3. A demonstration where I take their actual assignments and complete them in 30 seconds using generative AI
  4. sadness.
  5. And then, hope. Hope when they realize that the only solution is good teaching.

That is in no way meant to reflect a statement about all faculty or teachers. I only really get the ones who care about teaching and their students.

Anyway… this is what I’ve been telling them.


Generative AI has already changed education. Whether we realise it or not, every student has been forced to make a decision about whether they are going to use AI to generate text, videos or images. Of particular importance to those of us in the teaching profession, we have lost our ability to make students write their own texts when we are not watching them. Regardless of your position on the inclusion of generative AI in the classroom, this is likely to have a profound impact on your classroom. 

The following checklist is emergent. We will continue to add to it as situations develop and new approaches emerge. 

Dave Cormier and the Office of Open Learning, UWindsor.

Have I included my own stance regarding Generative AI in my syllabus?

There is no agreed upon way for people to be handing these systems right now. So, whether you’re talking about Chegg or ChatGPT, be clear in your syllabus how you expect students to use it (or not). This will not, on its own, stop people from doing things, but it will at least make it clear for someone who wants to do the right thing. 

  1. Tell them what you would like them to do
  2. Give them tools that you think are appropriate
  3. Find ways to incorporate ethical usage of these tools into your classroom teaching practice


Have I explained what counts as engagement in my course and explained why I want students to do the work that I am asking them to do? Have I added an explanation to each of my readings/assessments explaining to students why it’s important and how it is connected to the learning outcomes?

One of the ways of addressing students’ illegitimate usage of Generative AI tools is to explain to them why you want them to do the work laid out in your syllabus. We have, historically, forced students to do their homework by awarding them grades for completion. If students are doing their work to ‘complete’ it, instead of being driven by actual interest, they are going to be far more likely to find ways to complete their work without having to learn anything.

  1. Encourage students to find ways to be interested in the work in the classroom
  2. Share your own reasons for finding the material interesting
  3. Find ways to highlight examples of students performing in an engaged manner


Have I considered why I was using the take home assessments affected by generative AI (e.g. right answer questions, long form text like essays)? Can I replace them with new approaches that serve the same purpose?

There are many good arguments out there for the writing of essays and other long-form writing tasks. They promote deeper thinking, give students more space to construct critical arguments, and have strong disciplinary connections in some cases. They are a common means of demonstrating the student’s developing set of  research skills. In the past, we were often able to assume with reasonable confidence that essays and extended writing were a sound way to be assured that students were developing those skills. It has always been possible to pay someone else to write your essay, though and with the new tools available, and the relatively inexpensive rates available at essay mills, there is no longer any guarantee that any student is doing these things. 

  1. Consider using point form responses submitted in class
  2. Consider deconstructing the essay (an argument assignment, a research assignment) that never leads to a completed essay


Are there places where I am trying to ‘cover’ content that I can remove and allow for deeper work on more important topics?

There are many reasons that can lead to us needing to ‘cover’ certain kinds of material in a classroom. It could be that we are mandated by accreditation bodies, it could be that there is a course deeper in the degree that is counting on us to develop some foundational knowledge or skills. But this is not always the case. Many of us inherit the courses we teach and don’t entirely understand why a given topic is included in the course. Teaching less, and teaching more deliberately, allows us the time to delve into the nuances of a topic.


Have I provided enough time to allow my students to unlearn old ways of doing things before they can take on the new ones that I’m presenting?

The abilities that come with generative AI will likely lead to some changes in your course. It is critical that students get time to learn these new processes, so that they know what it means to be successful in your course. Skills that may have been valued when they were in high school may no longer be as important, and the changes made by one faculty member may not work in your class. Give students time to make those adaptations so they have their best chance at success.


Have I made ethical decisions about student data such that the assignments and activities in my class don’t require students to give away their own personal identification or their work to outside companies?

Each of the digital tools that we use in our classroom take different amounts of personal information and intellectual property from students. We can inquire of our IT departments for information regarding the usage of student data in our institutions. The guideline is simple: treat our student’s data the way we want our own data treated.


Have I reviewed how my field is being affected by the web and AI generation? If it’s significant, have I included this in my course?

Generative AI is going to have vastly different impacts by field. Reach out to your colleagues across your respective fields and get a sense of how AI is impacting their day-to-day work. Many disciplines have started to collect and share ideas within their communities of practice. Some professional associations have also provided their guidance. There is much still to learn as the use and capabilities of these tools evolves. 


Have I incorporated new best practices for finding/evaluating knowledge in my field that take AI generated content/marketed content into account? (e.g. Prompt engineering exercises)

As our knowledge work is increasingly mediated by algorithms like search engines and text generators, it’s vital that we learn how to best find, sort and evaluate information from these systems. While there are certainly common good practice approaches that apply across fields, some approaches (e.g. using curated databases) are going to be discipline specific. Incorporating activities that help learners use, manipulate and trick the algorithms to bring back the results they need, and of which they can be confident of reliability, are essential to developing 21st century literacies.

Given the abundance of information available, good and bad, often the most important literacy any student needs is to be able to sift through information to find what is most true or most useful. Building those activities into our courses might be the most important thing we can do to help students in their futures.


Have I confirmed that the changes I’ve made to my syllabus have not created an unfair amount of work for me or my students?

Anytime we rework our syllabus, there is a chance that we add more work than we had in our previous versions, sometimes without noticing. Make sure to consider the total number of hours that all of the planned activities (class time, labs, assignments, group work, independent research, reading, watching content, quizzes etc.) in our syllabus imposes on our students, and be careful to explain your work expectations to your students. Be mindful that students have many other classes, often with the same requirements as yours, as well as commitments outside of university. The more we load students with busy work, the less time they have to do the things that most of us value most – deeply engaging with the topics of our courses and demonstrating that engagement through our assessment tasks. 


Have I considered the accessibility implications of the digital tools I am using? Do they have the potential to improve or reduce accessibility?

Every tool comes with its own affordances. Think your way through the classroom advantages and disadvantages to any tool you are going to use. Has the tool been formally assessed for accessibility by the University? Have you talked to OOL or CTL about it? Have you tried checking it with an online accessibility checker such as WAVE or AccessiBe? Does it require new computers to run it? Does it require a login? Does it have visual elements that disadvantage some students? Do all images used have alternative text? Does it require high speed internet? Does it work the same way on a mobile device such as a phone or tablet? How does it interact with screen readers or text to speech tools? Does it require high levels of physical dexterity? Can it be controlled from a keyboard only? What is the cost of the tool and who is expected to pay for it? These and many more questions should inform any decision to use a technology in our teaching. 


Universal Design for Learning (UDL) for Inclusion, Diversity, Equity, and Accessibility (IDEA)

Getting Started With Accessibility

Creative Commons License
Except where otherwise noted, the content on this site is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.